In the area of medical stimulators, there is a trend towards an increased number of stimulation electrode sites to improve therapeutic efficacy by accurate stimulation of the intended target volume using field steering. Besides stimulation, there is an increased demand for accurate sensing of neural activity. Both trends require the presence of a relatively large cross-point switch matrix to couple stimulation and/or sensing electronics to selected probe electrode sites. The available volume for energy storage is decreasing in the state-of-the-art medical stimulators, although the required energy for brain stimulation is substantially constant. Consequently, there is less room for a battery, and, thus, the circuitry of the medical stimulator has to be a low power circuitry. The high number of switches of a cross-point switch matrix imposes an extremely-low power consumption requirement on a single switch with its control electronics.
The low-power requirement calls for integrated CMOS switches in a high-voltage IC technology, offering isolated NMOS and PMOS transistors. In state-of-the-art high-voltage IC technologies, the driving voltage of CMOS switches—the gate-to-source voltage—is limited to a few volts in comparison to the much higher voltage that is allowed across the CMOS switch itself—the drain-to-source voltage.
The article of W. N Reining, “A High voltage cross-point switch for medical applications”, Digest of the 1999 IEEE Southwest Symposium on Mixed-Signal Design SSMSD '99, Tucson, Ariz., USA, Apr. 11-13, 1999, pp. 109-112, discloses in FIG. 2 a bidirectional switch and a control circuit for the bidirectional switch for medical applications, such as medical stimulators. Two NMOS transistors M10, M11 of which the gates and the sources are coupled to each other form the bidirectional switch.
A current source, built with a high-voltage PMOS transistor M2, is connected between the common gate of the bidirectional switch transistor and a voltage supply terminal VHI which receives a voltage that is higher than ever is appearing at the bidirectional switch I/O terminals. To turn the bidirectional switch on, the current source M2 is conducting a small current, according to the article 3 μA. The current is conducted by a string of diode-connected NMOS transistors M4, M5, M6 and a high voltage PMOS transistor M9. The gate of M9 is connected to the common source of the bidirectional switch and the drain is connected to a voltage supply terminal VSS which receives a voltage that is at a voltage lower than ever is appearing at the I/O terminals of the bidirectional switch. The voltage drops across the forward-biased diode-connected transistors M4, M5 and M6 and the gate-source voltage of M9, several volts, switch the bidirectional switch to the on-state. It is to be noted that, when the bidirectional switch is in the on-state, the circuit dissipates an amount of power which is the product of the value of the current times the voltage difference between the voltages on the terminals VSS and VHI.
A second current source is built with high-voltage NMOS transistor M8 and is connected between the common gate of the bidirectional switch transistors M10 and M11 and the voltage supply terminal VSS. To control the bidirectional switch to be in the off state, the current source built with M8 is conducting a small current, which is also conducted through high voltage NMOS transistor M3. The gate of M3 is connected to the common source of the bidirectional switch and the drain is connected to the voltage supply terminal VHI. The voltage drop between the gate and the source of M3 switches the bidirectional switch in the off state. If the bidirectional switch is in the off state, an amount of power is dissipated that equals the value of the current times the voltage difference between the voltages on the terminals VSS and VHI.
Thus, the control circuit of the bidirectional switch of the cited articles has a static power dissipation and the dissipation is irrespective of the state of the bidirectional switch.